Game performance is an indicator of single-core performance, which is still very relevant for everything from web browsing to largely single-threaded applications like Blender.
Also, the best-selling and best-grossing CPUs are high-clocked quadcores, so it’s an important market that Zen cannot (yet) compete in.
I find that it is comparable to a i7-6900K and only using 90w most interesting.
It’s behind the i7-6900K by a margin of 10% (in either tests) when using 93W instead of 96W.
You guys should stop posting results rendered with the GPU’s cuz thats cheating.
And the Ryzen CPU was running at 3.4 GHz, overclocked surely there will be better results but lets stay modest here. If it matches a somewhat Intel CPU with 8 core 16 threads, we are going to see some competition again. Which means lower prices for us ^^
Dont hate it down guys
Well, i will not call it cheating, but it is out of purpose when looking at trying to see how good or bad is the Zen (Ryzen )results. Everyone now GPU’s are faster than any CPU’s for raytracing, its why we use them, its why there’s CUDA or OpenCL.
If it is for information purpose, it is good too to point for gpu’s performance, but can just put a link to the thread of GPU’s benchmark.
ooh, i will like a 16GB 1080ti, but well they need to have a coherent lineup for place price and performance point. ( and continue too, to sell costly workstation gpu’s ( Quadro or FirePro ).
But well, for professional of the industry you have now AMD who release gpu’s with embedded Nvme storage ( as high as 1To of storage ), and i know for sure that Nvidia will release the same things in his QUadro lineup later in 2017 -2018. ) For animation, VFX thats start to be a really nice addition as you can forget the Vram limitation ( for a minimal cost of performance if softwares are optimized for it )
Sry no, If you guys want to compare GPU Performance go into the GPU thread and compare Apples to apples. Some posts feel like nukes to Ryzen with saying “Oh the CPU isnt that great, look my 5 Titan X perform so much better!”. You wont compare a Car against a Jet after all, because obviously the Jet is a TON faster. Seriously Guys.
This has become a pissing contest. CPU’s can really double in speed per core. The only thing that can make things twice as fast are twice the cores.
Right now it’s about the (price & wattage)/calculation.
Those are engineering sample CPU’s, I doubt Intel has made them for selling at that price. A legal, retail eqivelent of that CPU cost > $3000 AUD, or at least > $2000 USD.
I don’t know. Those are engineer examples.
All I can see is a trade amount of more than 50000 Intel ES CPUs from different shops since my first attention to this kind of CPU around 2013.
My Xeon E5 V3 and V4 ES CPUs are already over 1 year old, you can buy Xeon E5 V5 ES for very high prices already, I am sure they will be available on Ebay soon.
They are normally first sold/seen few months before any official release, i.e. there aren’t any motherboard for the Xeon E5 V5 (Skylake-EP) available yet, buying them now won’t make sense unless you have connections to Taiwanese motherboard manufacturers.
I got 33.29 sec. with a slightly overclocked 5820K @3.8 Ghz and stock Blender 2.78a on Linux Mint 18, I didn’t change anything else, just loaded the .blend file.
From my perspective, using a non standard build to compare is not a fair approach. If they used the standard build, it would be good to see results from that.
We all know that performance from build to build can vary significantly, so unfair comparison.
As for Engineering Samples, that is one cool find.
@ Grimm, nice results. will have to find an Linux boot disk with Blender to do a quick test and see the render speed differences between Windows 10 and Linux. I know it’s there but wonder how much. My stock 2x e5-2687w give me 26s with default Blender. I so need to give a Linux a try…